remanso-cli / packages / cli / src / lib / markdown.ts

import { webcrypto as crypto } from "node:crypto";
import * as fs from "node:fs/promises";
import * as path from "node:path";
import { glob } from "glob";
import yaml from "js-yaml";
import { minimatch } from "minimatch";
import * as toml from "smol-toml";
import type { BlogPost, FrontmatterMapping, PostFrontmatter } from "./types";

export function parseFrontmatter(
	content: string,
	mapping?: FrontmatterMapping,
): {
	frontmatter: PostFrontmatter;
	body: string;
	rawFrontmatter: Record<string, unknown>;
} {
	// Support multiple frontmatter delimiters:
	// --- (YAML) - Jekyll, Astro, most SSGs
	// +++ (TOML) - Hugo
	// *** - Alternative format
	const frontmatterRegex = /^(---|\+\+\+|\*\*\*)\n([\s\S]*?)\n\1\n([\s\S]*)$/;
	const match = content.match(frontmatterRegex);

	if (!match) {
		const [, titleMatch] = content.trim().match(/^# (.+)$/m) || [];
		const title = titleMatch ?? "";
		const [publishDate] = new Date().toISOString().split("T");

		return {
			frontmatter: {
				title,
				publishDate: publishDate ?? "",
			},
			body: content,
			rawFrontmatter: {
				title: publishDate,
			},
		};
	}

	const delimiter = match[1];
	const frontmatterStr = match[2] ?? "";
	const body = match[3] ?? "";

	// Determine format based on delimiter:
	// +++ uses TOML (key = value)
	// --- and *** use YAML (key: value)
	const isToml = delimiter === "+++";

	// Parse frontmatter using the appropriate library
	let raw: Record<string, unknown>;
	if (isToml) {
		raw = toml.parse(frontmatterStr) as Record<string, unknown>;
	} else {
		// Use CORE_SCHEMA to keep dates as strings rather than Date objects
		raw =
			(yaml.load(frontmatterStr, { schema: yaml.CORE_SCHEMA }) as Record<
				string,
				unknown
			>) ?? {};
	}

	// Apply field mappings to normalize to standard PostFrontmatter fields
	const frontmatter: Record<string, unknown> = {};

	// Title mapping
	const titleField = mapping?.title || "title";
	frontmatter.title = raw[titleField] || raw.title;

	// Description mapping
	const descField = mapping?.description || "description";
	frontmatter.description = raw[descField] || raw.description;

	// Publish date mapping - check custom field first, then fallbacks
	const dateField = mapping?.publishDate;
	if (dateField && raw[dateField]) {
		frontmatter.publishDate = raw[dateField];
	} else if (raw.publishDate) {
		frontmatter.publishDate = raw.publishDate;
	} else {
		// Fallback to common date field names
		const dateFields = ["pubDate", "date", "createdAt", "created_at"];
		for (const field of dateFields) {
			if (raw[field]) {
				frontmatter.publishDate = raw[field];
				break;
			}
		}
	}

	// Cover image mapping
	const coverField = mapping?.coverImage || "ogImage";
	frontmatter.ogImage = raw[coverField] || raw.ogImage;

	// Theme, font family and font size
	frontmatter.theme = raw.theme;
	frontmatter.fontFamily = raw.fontFamily;
	frontmatter.fontSize = raw.fontSize;

	// Tags mapping
	const tagsField = mapping?.tags || "tags";
	frontmatter.tags = raw[tagsField] || raw.tags;

	// Draft mapping
	const draftField = mapping?.draft || "draft";
	const draftValue = raw[draftField] ?? raw.draft;
	if (draftValue !== undefined) {
		frontmatter.draft = draftValue === true || draftValue === "true";
	}

	// Discoverable mapping (defaults to true if not set or invalid)
	const discoverableValue = raw.discoverable;
	if (discoverableValue === false || discoverableValue === "false") {
		frontmatter.discoverable = false;
	} else {
		frontmatter.discoverable = true;
	}

	// Always preserve atUri (internal field)
	frontmatter.atUri = raw.atUri;

	return {
		frontmatter: frontmatter as unknown as PostFrontmatter,
		body,
		rawFrontmatter: raw,
	};
}

export function getSlugFromFilename(filename: string): string {
	return filename
		.replace(/\.mdx?$/, "")
		.toLowerCase()
		.replace(/\s+/g, "-");
}

export interface SlugOptions {
	slugField?: string;
	removeIndexFromSlug?: boolean;
	stripDatePrefix?: boolean;
}

export function getSlugFromOptions(
	relativePath: string,
	rawFrontmatter: Record<string, unknown>,
	options: SlugOptions = {},
): string {
	const {
		slugField,
		removeIndexFromSlug = false,
		stripDatePrefix = false,
	} = options;

	let slug: string;

	// If slugField is set, try to get the value from frontmatter
	if (slugField) {
		const frontmatterValue = rawFrontmatter[slugField];
		if (frontmatterValue && typeof frontmatterValue === "string") {
			// Remove leading slash if present
			slug = frontmatterValue
				.replace(/^\//, "")
				.toLowerCase()
				.replace(/\s+/g, "-");
		} else {
			// Fallback to filepath if frontmatter field not found
			slug = relativePath
				.replace(/\.mdx?$/, "")
				.toLowerCase()
				.replace(/\s+/g, "-");
		}
	} else {
		// Default: use filepath
		slug = relativePath
			.replace(/\.mdx?$/, "")
			.toLowerCase()
			.replace(/\s+/g, "-");
	}

	// Remove /index or /_index suffix if configured
	if (removeIndexFromSlug) {
		slug = slug.replace(/\/_?index$/, "");
	}

	// Strip Jekyll-style date prefix (YYYY-MM-DD-) from filename
	if (stripDatePrefix) {
		slug = slug.replace(/(^|\/)(\d{4}-\d{2}-\d{2})-/g, "$1");
	}

	return slug;
}

export function slugifyTitle(title: string): string {
	return (title || "")
		.toLowerCase()
		.replace(/\s+/g, "-")
		.replace(/[^\w-]/g, "");
}

export async function getContentHash(content: string): Promise<string> {
	const encoder = new TextEncoder();
	const data = encoder.encode(content);
	const hashBuffer = await crypto.subtle.digest("SHA-256", data);
	const hashArray = Array.from(new Uint8Array(hashBuffer));
	return hashArray.map((b) => b.toString(16).padStart(2, "0")).join("");
}

function shouldIgnore(relativePath: string, ignorePatterns: string[]): boolean {
	for (const pattern of ignorePatterns) {
		if (minimatch(relativePath, pattern)) {
			return true;
		}
	}
	return false;
}

export interface ScanOptions {
	frontmatterMapping?: FrontmatterMapping;
	ignorePatterns?: string[];
	slugField?: string;
	removeIndexFromSlug?: boolean;
	stripDatePrefix?: boolean;
}

export async function scanContentDirectory(
	contentDir: string,
	frontmatterMappingOrOptions?: FrontmatterMapping | ScanOptions,
	ignorePatterns: string[] = [],
): Promise<BlogPost[]> {
	// Handle both old signature (frontmatterMapping, ignorePatterns) and new signature (options)
	let options: ScanOptions;
	if (
		frontmatterMappingOrOptions &&
		("frontmatterMapping" in frontmatterMappingOrOptions ||
			"ignorePatterns" in frontmatterMappingOrOptions ||
			"slugField" in frontmatterMappingOrOptions)
	) {
		options = frontmatterMappingOrOptions as ScanOptions;
	} else {
		// Old signature: (contentDir, frontmatterMapping?, ignorePatterns?)
		options = {
			frontmatterMapping: frontmatterMappingOrOptions as
				| FrontmatterMapping
				| undefined,
			ignorePatterns,
		};
	}

	const {
		frontmatterMapping,
		ignorePatterns: ignore = [],
		slugField,
		removeIndexFromSlug,
		stripDatePrefix,
	} = options;

	const patterns = ["**/*.md", "**/*.mdx"];
	const posts: BlogPost[] = [];

	for (const pattern of patterns) {
		const files = await glob(pattern, {
			cwd: contentDir,
			absolute: false,
		});

		for (const relativePath of files) {
			// Skip files matching ignore patterns
			if (shouldIgnore(relativePath, ignore)) {
				continue;
			}

			const filePath = path.join(contentDir, relativePath);
			const rawContent = await fs.readFile(filePath, "utf-8");

			try {
				const { frontmatter, body, rawFrontmatter } = parseFrontmatter(
					rawContent,
					frontmatterMapping,
				);
				const slug = getSlugFromOptions(relativePath, rawFrontmatter, {
					slugField,
					removeIndexFromSlug,
					stripDatePrefix,
				});

				posts.push({
					filePath,
					slug,
					frontmatter,
					content: body,
					rawContent,
					rawFrontmatter,
				});
			} catch (error) {
				console.error(`Error parsing ${relativePath}:`, error);
			}
		}
	}

	// Sort by publish date (newest first)
	posts.sort((a, b) => {
		const dateA = new Date(a.frontmatter.publishDate);
		const dateB = new Date(b.frontmatter.publishDate);
		return dateB.getTime() - dateA.getTime();
	});

	return posts;
}

export function updateFrontmatterWithAtUri(
	rawContent: string,
	atUri: string,
): string {
	// Detect which delimiter is used (---, +++, or ***)
	const delimiterMatch = rawContent.match(/^(---|\+\+\+|\*\*\*)/);
	const delimiter = delimiterMatch?.[1] ?? "---";
	const isToml = delimiter === "+++";

	// Format the atUri entry based on frontmatter type
	const atUriEntry = isToml ? `atUri = "${atUri}"` : `atUri: "${atUri}"`;

	// No frontmatter: create one with atUri
	if (!delimiterMatch) {
		return `---\n${atUriEntry}\n---\n\n${rawContent}`;
	}

	// Check if atUri already exists in frontmatter (handle both formats)
	if (rawContent.includes("atUri:") || rawContent.includes("atUri =")) {
		// Replace existing atUri (match both YAML and TOML formats)
		return rawContent.replace(
			/atUri\s*[=:]\s*["']?[^"'\n]+["']?\n?/,
			`${atUriEntry}\n`,
		);
	}

	// Insert atUri before the closing delimiter
	const frontmatterEndIndex = rawContent.indexOf(delimiter, 4);
	if (frontmatterEndIndex === -1) {
		throw new Error("Could not find frontmatter end");
	}

	const beforeEnd = rawContent.slice(0, frontmatterEndIndex);
	const afterEnd = rawContent.slice(frontmatterEndIndex);

	return `${beforeEnd}${atUriEntry}\n${afterEnd}`;
}

export function removeFrontmatterAtUri(rawContent: string): string {
	const frontmatterRegex = /^(---|\+\+\+|\*\*\*)\n([\s\S]*?)\n\1\n/;
	const match = rawContent.match(frontmatterRegex);
	if (!match) return rawContent;

	const delimiter = match[1];
	const frontmatterStr = match[2] ?? "";

	// Remove the atUri line
	const lines = frontmatterStr
		.split("\n")
		.filter((line) => !line.match(/^\s*atUri\s*[=:]\s*/));

	// Check if remaining frontmatter has any non-empty lines
	const hasContent = lines.some((line) => line.trim() !== "");

	const afterFrontmatter = rawContent.slice(match[0].length);

	if (!hasContent) {
		// Remove entire frontmatter block, trim leading newlines
		return afterFrontmatter.replace(/^\n+/, "");
	}

	return `${delimiter}\n${lines.join("\n")}\n${delimiter}\n${afterFrontmatter}`;
}

export function stripMarkdownForText(markdown: string): string {
	return markdown
		.replace(/#{1,6}\s/g, "") // Remove headers
		.replace(/\*\*([^*]+)\*\*/g, "$1") // Remove bold
		.replace(/\*([^*]+)\*/g, "$1") // Remove italic
		.replace(/\[([^\]]+)\]\([^)]+\)/g, "$1") // Remove links, keep text
		.replace(/`{3}[\s\S]*?`{3}/g, "") // Remove code blocks
		.replace(/`([^`]+)`/g, "$1") // Remove inline code formatting
		.replace(/!\[.*?\]\(.*?\)/g, "") // Remove images
		.replace(/\n{3,}/g, "\n\n") // Normalize multiple newlines
		.trim();
}

export function getTextContent(
	post: { content: string; rawFrontmatter?: Record<string, unknown> },
	textContentField?: string,
): string {
	if (textContentField && post.rawFrontmatter?.[textContentField]) {
		return String(post.rawFrontmatter[textContentField]);
	}
	return stripMarkdownForText(post.content);
}